Skip to main content
GET
Universal Search

​
Filters

Universal Search accepts an optional filters query parameter. The parameter is parsed as JSON, so you can either send a JSON object or a JSON string (for example, filters={"blockchains":"1,solana"} or its URL-encoded form). Keys and values are case-insensitive and surrounding whitespace is ignored.

​
blockchains

  • Accepts either a comma-separated string or a JSON array.
  • Supports both numeric chain identifiers (for example 1, 8453, 42161) and string identifiers (for example solana, evm:1, tron:sol).
  • You can freely mix numeric and string identifiers in the same request; the backend normalizes them before querying.
  • Passing multiple IDs limits the search results to tokens that exist on any of the provided chains.
Example request
This query searches for eth while restricting the results to Base (8453), Ethereum (1), Arbitrum (42161), Optimism (10), and Solana.

​
Search Modes

​
Prefix Search (default)

By default, the search performs a prefix match on token names and symbols. For example, searching W will return tokens like “WETH”, “WBTC”, “Wrapped SOL”, etc. For tokens whose on-chain name or symbol is not latin-script (e.g. Chinese meme tokens), the search also matches their stored translations: searching Niu finds 牛来 (translated “Niu Lai”), and searching 龙 finds tokens whose Chinese translation starts with 龙. If input starts with an EVM address prefix such as 0x0e09, Universal Search also matches token contract addresses, pool addresses, and deployer addresses that start with that prefix.
This is useful for address autocomplete flows where a user pastes only the first characters of a contract address.

​
Exact Match (advanced)

Wrap your search term in double quotes to perform an exact match on the symbol. This is useful when you want to find a specific token without getting results that merely start with your query. Example: "W" will only return tokens with the exact symbol “W” (like Wormhole), not “WETH” or “WBTC”. Exact match is literal-only: translated names are not considered in this mode.
Use exact match when searching for short symbols (1-3 characters) to avoid noisy results.
You can sort by an array of metrics using sortBy (default: searchScore):
  • volume24h
  • marketCap
  • createdAt
  • volume1h
  • feesPaid5min
  • feesPaid1h
  • feesPaid24h
  • volume5min
  • holdersCount
  • organicVolume1h
  • totalFeesPaidUsd
  • searchScore
  • liquidity
searchScore is a composite ranking metric calculated as fees_paid_24h × quote_token_reserve_usd (using the largest pool by liquidity). This metric prioritizes tokens with both high trading fees and deep liquidity, making it ideal for discovering actively traded tokens with strong market depth.
You can control the number of results using the limit parameter (1 to 20, default: 5).
Performance scales with the number of results requested. Expect ~50ms for 5 results and up to ~250ms for 20 results.
Alpha Feature: When searching by address, the search also returns all tokens deployed by that address (deployer match).

​
Cookbook: multi-chain filtering

  1. Identify the chain identifiers you want to include. You can use numeric IDs (1, 56, 8453) or string-based IDs (solana, evm:1, evm:8453). Both styles are supported in the same request.
  2. Build the JSON payload for filters. Set the blockchains property to a comma-separated string of the identifiers (e.g. "blockchains":"8453,1,solana"). You may also send an explicit JSON array such as "blockchains":["8453","1","solana"].
  3. URL-encode the JSON when embedding it inside a query string. Most HTTP libraries will do this automatically if you pass an object.
  4. Call /api/2/fast-search?input=<term>&filters=<encoded-json>. The API returns only the top matches that exist on any of the requested chains, making it straightforward to focus on supported networks without enabling all chains globally.
Tip: pair this recipe with the limit and sortBy parameters to surface the most liquid matches on the networks that matter to you.

🔀 Need AND / OR / NOT filters?

The POST variant accepts a JSON body with the same payload shape as a single pulse v2 view — top-level chainId / poolTypes / sortBy / limit, plus a Prisma-like filters tree with AND / OR / NOT and the full leaf-operator set (gte, lte, in, contains, startsWith, …). Useful for per-chain conditional constraints that can’t be expressed as flat GET filters.

Query Parameters

​
input
string
required

Search query string

​
type
enum<string>

Type of results to return

Available options:
tokens,
assets,
pairs
​
filters
string

JSON string with filter options: blockchains, poolTypes, excludeBonded, bondedOnly

​
sortBy
enum<string>

Sort field for search results

Available options:
volume24h,
marketCap,
createdAt,
volume1h,
feesPaid5min,
feesPaid1h,
feesPaid24h,
volume5min,
holdersCount,
organicVolume1h,
totalFeesPaidUsd,
searchScore,
liquidity
​
excludeBonded
boolean

Exclude bonded tokens from results

​
limit
number

Maximum number of results (1-20, default: 5)

​
categories
string[]

Asset category names to filter on, backed by /api/1/metadata/categories

Response

200 - application/json

Universal Search response

​
data
object[]
required